Lane Kenworthy, The Good Society
April 2023

Democracy is a system of decision making in which participants have approximately equal opportunity to influence policy choices.1 This entails:

  1. Each person has the same number of votes in electing policy makers (representative democracy) and in direct policy making (direct democracy).
  2. Each person has roughly the same opportunity to influence policy makers’ views and actions via organization, lobbying, protest, and other activities.
  3. Individuals have access to adequate information in order to develop informed preferences.
  4. Decisions are made according to majority rule (though the majority can’t abridge conditions 1-3).

Democracy is relatively new. As figure 1 shows, two centuries ago hardly anyone lived in a country with a democratic political system. But since then democracy has been on the rise.

Figure 1. Democracy
Share of the world’s population living in a country with a democratic political system. Polity IV democracy scoring ranges from -10 for a hereditary monarchy to +10 for a consolidated democracy. Data source: Max Roser and Bastian Herre, “World Citizens Living Under Different Political Regimes,” Our World in Data.

The United States has the world’s longest-running political democracy. One of its most remarkable features, highlighted in figure 2, is that for more than two centuries the losing candidates in elections to the presidency, the House of Representatives, and the Senate have peacefully allowed the winning candidates to take office. The government and political system have survived potentially mortal threats, most notably the southern states’ attempted secession and subsequent civil war in the 1860s. Voting rights have been steadily expanded over time — to non-property-owners by the 1830s, to African Americans formally in 1870 and substantively a century later, to women in 1920, to 18-to-20-year-olds in 1971. The system’s stability and popular legitimacy have enabled reasonably wise policy making and helped to facilitate a massive rise in living standards for America’s citizens.

Figure 2. Elections with peaceful transfer of power
Share of elections. 1789-present. The only exception was the 1860 presidential election: after Abraham Lincoln was elected president in November 1860, seven southern states declared secession from the country before Lincoln took office in March of 1861.

While deserving of praise, America’s democracy is far from perfect. It may discourage voting and other forms of political activity (speech, organization, lobbying, etc.) among Americans with less education and income. The electoral systems for the Senate, the House of Representatives, and the presidency translate votes into representation less effectively than we might wish. Private money may give businesses and wealthy individuals disproportionate influence on policy making. Our veto-point-heavy government structure may go too far in preserving policy continuity at the expense of the majority’s will.

To what extent do these problems exist? Have they worsened over time?

Skip to:


Does our voting system result in more high-income voters than low-income voters?

As figure 3 shows, in recent decades only about half of voting-age Americans have tended to vote in presidential-year elections. In “off-year” elections, only a third have. These shares have been lower than in many other rich longstanding-democratic nations.

Figure 3. Voter turnout
Share of the voting-age population. Legislative elections. The solid thick line is for US presidential election years; the dashed thick line is for US off-years. Data source: International Institute for Democracy and Electoral Assistance,, series VAP turnout. “Asl” is Australia; “Aus” is Austria.

Figure 4 shows voter turnout by socioeconomic status (SES, a measure combining education and income). The higher the socioeconomic status, the more likely the person is to vote.

Figure 4. Voter turnout by socioeconomic status
Share of eligible voters. “SES” = socioeconomic status, a measure that combines education and income. Q1 is the lowest 20% on the socioeconomic status ladder; Q5 is the highest. Data source: Kay Lehman Schlozman, Henry E. Brady, and Sidney Verba, Unequal and Unrepresented, Princeton University Press, 2018, figure 11.2, using data from the American National Elections Studies (ANES). These survey data overestimate the level of voter turnout by about 15 percentage points, but they are likely to accurately reflect SES differences in voting and trends over time.

The class skew in voter turnout owes partly to preferences, but there are systemic obstacles in the US that also matter. Only two-thirds of eligible Americans are registered to vote, because in many states registration must be done in advance and in person. Voting too usually is carried out in person, and on a work day during working hours. Some voting-age adults are barred from voting: noncitizens, prisoners, ex-felons, persons who can’t produce “proper” identification. These obstacles reduce voting more among Americans with lower eduction and/or income. We know that because the class skew in voting widened in the United States when registration requirements were introduced after 1896 and because the skew is larger in the US than in rich countries that have different registration and voting procedures.2

In recent years, as opportunities to vote early and by mail have expanded, turnout has increased significantly (figure 3).3 We don’t yet know whether, and if so how much, this has reduced the class skew in voting.


An electoral system that effectively translates votes into representation is needed to ensure that each person’s vote has the same weight. America’s electoral system has several problematic features.

Most rich longstanding-democratic countries have a proportional representation electoral system, in which voting is (solely or mainly) for parties and each party’s share of seats in the legislature corresponds to the share of votes it receives. The United States has winner-take-all elections. Votes are cast for candidates, and the candidate receiving the most votes in a district is the only representative of that district. Winner-take-all systems tend to underrepresent minority views and preferences, as smaller parties, the most likely source of such representation, seldom gain political traction. It’s difficult for them to win elections, so their support fades over time and they disappear. Over the past 150 years, no third party has been able to survive in American politics at the national level.4

The Senate is one of the most malapportioned legislative bodies in the world, with Americans who live in low-population states far overrepresented relative to their counterparts in high-population states.5 As figure 5 shows, the 22 least-populated states have a total population equal to that of California. The 39 million Americans in those 22 states are represented by 44 senators, while the 39 million Americans in California are represented by two.

Figure 5. Senate malapportionment
“m” = million. Data source: Census Bureau. A different version of this chart appears in Ezra Klein, “If You’re from California, You Should Hate the Senate,” Washington Post: Wonkblog, 2013.

In 2017 and 2018, for example, Democratic Party senators represented 55% of the US population, Republican Party senators 45% (if we treat each senator as representing half of the population of her or his state). Yet the Democrats had only 48 of the senators, the Republicans 52.6 This type of outcome, with the party that got fewer votes ending up with a majority of Senate seats, has occurred at least seven other times in the past century.7

Unlike the Senate, the House of Representatives has districts of (approximately) equal population size. Yet it too fails to effectively translate party vote share into party representation share. That’s plainly true for third parties, which seldom get any representation regardless of whether they get 2% of the vote or 20% of the vote. But it’s even true for the two major parties. In the 2016 House elections, for example, Republican candidates got 49.6% of the popular vote but ended up with 55% of the seats.8

One reason for this is the heavy concentration of Democratic voters in large cities and hence within certain House districts. This results in more Democratic votes being “wasted” than Republican votes: more Democratic candidates win by a margin of, say, 80% to 20%, whereas Republican winners are more likely do so by a margin of 55% to 45%. Democrats thereby win fewer seats than if their voters were more spread out across House districts.9

It’s also partly a result of the gerrymandering of House districts. Every ten years, following the decennial census, the districts are redrawn to respond to population shifts. In most states, the party that controls the state government at that moment gets to decide how to draw the new district boundaries. Partisan gerrymandering consists of creating a number of districts in which your party will win by a comfortable but not excessive margin and a few districts in which the other party will win with nearly all of the votes. This results in more of the other party’s votes being wasted, allowing your party to win a larger number of districts. Republicans were able to do this in more states than Democrats following the 2010 census.10

Figure 6 illustrates how gerrymandering can be done. Imagine a state with 50 precincts (or 50 voters), 60% of which are Democratic (blue) and 40% of which are Republican (red). The state has five House districts. Those five districts can be drawn in a variety of different ways, some of which accurately reflect the balance within the state as a whole (second column), others that give the Democrats more seats than they should have (third column), and still others that overrepresent the Republicans (fourth column).

Figure 6. Illustration of gerrymandering
Hypothetical state with five House of Representatives districts. Source: Christopher Ingraham, “Take a Look at One of the Country’s Most Blatant Gerrymanders,” Washington Post: Wonkblog, 2017.

A growing number of states have turned over the redrawing of House districts to independent commissions, and this has reduced gerrymandering following the 2020 Census. In the 2022 House elections, Republicans won 50.6% of the popular vote and ended up with 51.0% of the seats.

As with the Senate and the House of Representatives, there is a major flaw in how votes are translated into representation in US presidential elections. Rather than simply tallying up the popular vote, the presidential winner is decided via the electoral college. The candidate who wins the popular vote in each state gets all of that state’s electoral college votes, regardless of the margin of victory.11 When the country is evenly divided between the two parties, there is a significant possibility that the candidate who gets the most votes will nevertheless lose the electoral college.

That has happened only twice in the modern democratic era (since 1920, when women got the right to vote). But those two occurrences, in 2000 and 2016, were among the six most recent elections. It nearly happened also in 2004 and 2020. Had John Kerry gotten 120,000 more votes in Ohio in 2004, he, like George W. Bush and Donald Trump, would have become president despite losing the popular vote. And had Trump gotten 15,000 more votes in Arizona and Georgia and 25,000 more in Wisconsin in 2020, he would have twice in a row won the electoral college despite losing the popular vote.


The earliest systematic studies of Americans’ political beliefs and preferences, in the 1940s and 1950s, found that most citizens don’t pay close attention to most policy issues and so don’t have an opinion on them. This is one reason why only a modest share vote. And those who do vote, even those who are well informed and politically engaged, tend to develop preferences and cast votes based on their group identity, partisan loyalty, and the perceived condition of the economy, rather than informed reflection on the issues.12

This owed partly to lack of information. Circumstances today have changed dramatically. With the internet, most citizens have relatively easy access to more than enough information they need to develop informed preferences about issues, parties, and candidates. Yet recent studies, including a replication of the hugely influential The American Voter fifty years on, have found no improvement in citizens’ political awareness or ideological consistency. Political junkies are better informed than ever before, but the average American isn’t.13

One reason is that voter ignorance is a product not only of lack of information but also lack of time and interest. It’s also the case, though, that together with the massive increase in access to information has come a parallel increase in faux information and fake news — “facts” and “stories” that are inaccurate or intentionally misleading.


If voters have trouble forming preferences on multiple issues and matching those preferences to candidates, proponents of direct democracy contend that they can do better on individual issues. In about half of the US states, Americans regularly cast yes or no votes on an assortment of initiatives and referendums, making policy decisions directly rather than indirectly through elected representatives. Most of these procedures were created during the progressive era in the early twentieth century.

It isn’t clear whether this has improved our democracy. Some studies conclude that state initiatives and referendums have tended to produce better policy decisions than the regular legislative process.14 Others point out that those with lots of money have a greater advantage in swaying voters to pass a ballot initiative than in getting lawmakers to follow their wishes, that citizens frequently vote in favor of populist but shortsighted reforms such as term limits and tax cuts, and that the yes-or-no format of direct democracy weakens the quality of policy outputs.15


Voting and the translation of votes into representation are democracy’s core inputs. What about the outputs? To what extent do policy decisions reflect the will of the people — or, if not their will, perhaps their interests?16

One reason for skepticism is that, as with voting, participation in other political activities varies according to socioeconomic status. Wealthy business leaders serve as close advisors to presidents.17 Firms and groups that represent them (the Chambers of Commerce, National Association of Manufacturers, industry associations) spend large quantities of money lobbying policy makers. Rich Americans shower campaign contributions on their preferred candidates. The most thorough research on this issue, by Kay Lehman Schlozman, Sidney Verba, and Henry Brady, finds that these anecdotal impressions reflect the broader pattern. “The individuals and organizations that are active in American politics,” they write, “are anything but representative. In particular, those who are not affluent and well educated are less likely to take part politically.”18 As figure 7 shows, Schlozman and colleagues find that Americans with less education and income are less likely to contact a government official, work with fellow citizens to solve a community problem, make a monetary political contribution, work for a political party or candidate, or attend a protest.

Figure 7. Political activity (other than voting) by socioeconomic status
Share of adults engaging in at least one political act other than voting. “SES” = socioeconomic status, a measure that combines education and income. Q1 is the lowest 20% on the socioeconomic status ladder; Q5 is the highest. The data for 1967, 1990, and 2008 are based on the same five measures of activity: working in a political campaign, contributing to a candidate or campaign, contacting a government official, belonging to a political organization, and working with others on a community issue. The data for 1960 are based on a somewhat different set of activities, although ones that closely parallel those used in the other studies. Data source: Kay Lehman Schlozman, Henry E. Brady, and Sidney Verba, Unequal and Unrepresented, Princeton University Press, 2018, figure 11.1, using data from the Civic Culture Study 1960, Political Participation in America 1967, Citizen Participation Study 1990, and Pew Internet and American Life Survey 2008.

Political parties, advocacy groups, and social movements attempt to represent and promote citizens’ interests and preferences. Indeed, according to a distinguished line of political analysis, from E.E. Schattschneider to Thomas Ferguson and Joel Rogers to Jacob Hacker and Paul Pierson, government policy in the United States is determined mainly by the relative strength of organized interest groups.19 Does this help to compensate for the inequality in citizens’ political participation? Or does it accentuate that inequality?

Businesses and rich individuals have a lot of money, and both the absolute amount and their share have increased since the late 1970s. They’ve also organized into groups — business organizations, foundations, think tanks, political action committees, lobbying agencies — in order to increase their influence on policy making. During this same period, organizations representing ordinary Americans, most notably labor unions, have weakened. Unlike some other affluent democratic nations, the United States has no institutionalized mechanism whereby organizations representing both the more and less powerful discuss or negotiate important policy issues (sometimes called “corporatism”).20 Do affluent Americans and business firms exert disproportionate influence on policy decisions? Has their influence increased in recent decades?

Figure 8 shows the share of high-, middle-, and low-SES Americans who donate money to a political campaign. Figure 9 shows the share of campaign contributions that come from the top .01% of donors. Figure 10 shows the trend in total campaign contributions over the period for which we have reliable data, beginning in 1998. Figure 11 shows the trend in lobbying expenditures. Figure 12 shows the trend in labor union membership. Together, these data tell us that rich individuals and businesses are indeed the most likely to insert money into the political process, that the quantities they give to political campaigns and to lobbyists are large and have increased in recent decades, and that the organizational strength of the chief group representing less-educated and lower-income Americans has declined significantly.

Figure 8. Campaign contribution by socioeconomic status
Share of adults. “SES” = socioeconomic status, a measure that combines education and income. Q1 is the lowest 20% on the socioeconomic status ladder; Q5 is the highest. Data source: Kay Lehman Schlozman, Henry E. Brady, and Sidney Verba, Unequal and Unrepresented, 2018, figure 11.4, using data from the American National Elections Studies (ANES).

Figure 9. Campaign contributions by the top .01% of donors
Share of campaign contributions in all federal elections donated by the top .01% of the voting age population. The figure includes individual contributions to super PACS and 527 organizations but excludes contributions to nondisclosing 501c(4) organizations, which raise much of their money from wealth individuals. Including contributions to nondisclosing 501c(4)s might increase the share by 1-2 percentage points in 2010 and 2012. Data source: Adam Bonica, Nolan McCarty, Keith T. Poole, and Howard Rosenthal, “Why Hasn’t Democracy Slowed Rising Inequality?,” Journal of Economic Perspectives, 2014, figure 5.

Figure 10. Campaign expenditures
Billions of inflation-adjusted dollars. Includes expenditures by candidates, parties, and outside groups. Data source: Center for Responsive Politics, “Cost of Election,”

Figure 11. Lobbying expenditures
Billions of inflation-adjusted dollars. Data source: Center for Responsive Politics, “Lobbying Data Summary,”

Figure 12. Labor union membership
Share of employees who are union members. Data sources: 1933-82 are from Richard B. Freeman, “Spurts in Union Growth: Defining Moments and Social Processes,” in The Defining Moment: The Great Depression and the American Economy in the Twentieth Century, edited by Michael D. Bordo et al, University of Chicago Press, 1998, table 8A.2. 1983ff are from Bureau of Labor Statistics,, series LUU0204899600, using Current Population Survey data.

It also is true that Americans with higher incomes are more likely to get the policy they want than are those with middle and low incomes. In separate analyses, Larry Bartels, Martin Gilens, and Gilens and Benjamin Page have found that policy choices are much more likely to reflect the expressed policy preferences of high-income Americans than of those lower on the income ladder.21 Figure 13 depicts this finding graphically. The preferences of Americans near the top of the income distribution are positively correlated with policies adopted by congress and the president (solid line). The preferences of those with low incomes are, by contrast, uncorrelated with actual policy choices (dashed line).

Figure 13. Policy responsiveness by income
Policy responsiveness in instances where the preferences of high-income Americans differ from those of low-income Americans. “P10” is the tenth percentile of the income distribution; “p90” is the 90th percentile. This replicates figure 3.5 in Martin Gilens, Affluence and Influence, Princeton University Press, 2012.

What about the influence of interest groups? Here research conclusions are mixed. Some studies have documented a number of important instances of policy change and non-change in recent decades in which business firms and associations have gotten what they wanted.22 On the other hand, studies of lobbying have tended to find little or no systematic impact.23

Given the increase in income inequality in the United States since the late 1970s, we might expect the class skew in political participation to have increased in recent decades.24 Figure 7 (above) suggests, however, that that hasn’t happened.25 One reason might be that the rise in income inequality has consisted mainly of growing separation between the top 1% and everyone else, rather than between, say, the middle class and the poor.26

Has the disproportionate influence of the rich and/or the business community increased in recent decades? Studies that document policy wins by business and affluent Americans haven’t tended to find a rise in the frequency of such wins in recent decades.27 Martin Gilens has examined the correlation between income and influence on policy at various points over the past half century. As figure 14 shows, Gilens finds that the gap in influence between Americans with high incomes and those with low incomes was small during the Johnson presidency, larger during the presidencies of Reagan and Clinton, but then smaller during the first six years of George W. Bush’s presidency.

Figure 14. Policy responsiveness by income over time
Policy responsiveness in instances where the preferences of high-income Americans differ from those of low-income Americans. “P10” is the tenth percentile of the income distribution; “p90” is the 90th percentile. Data source: Martin Gilens, Affluence and Influence, Princeton University Press, 2012, figure 7.4.

If inequality of influence in fact hasn’t increased, it may be because money has diminishing returns in politics. It’s possible that we reached the saturation point decades ago, so that additional money no longer buys much, or perhaps any, additional influence.28


America’s founders structured the country’s government to safeguard against excessive power of the majority and against radical change.29 They did so by creating a large number of “veto points” — points at which proposed policy changes can be blocked:

  • A directly-elected executive (president). In most other rich longstanding-democratic countries, the executive (usually the prime minister) is the leader of the largest party in the parliament, and she or he has no independent lawmaking authority.
  • Two legislative bodies (the House of Representatives and Senate) with equal power. Other nations usually have just one house in parliament; where there are two, one tends to be largely ceremonial.
  • A judiciary with considerable leeway to void legislation by ruling it unconstitutional.
  • A highly federalized government, with state and local governments that have substantial decision-making power.

In addition, for the past century the Senate has operated under a self-imposed rule — the filibuster — that allows as few as 41 of its 100 members to block proposed policy changes. This means passing legislation in the Senate often requires a supermajority of 60 votes, rather than a simple majority of 51.30

With so many veto points, it is, arguably, too difficult for the majority to implement its policy preferences. Benjamin Page and Martin Gilens examined 1,791 policy changes proposed between 1981 and 2002. Figure 15 shows that when more Americans favor a proposed change, it is more likely to get adopted. Yet a relatively small share of proposals actually get passed. Even for proposed policy changes favored by virtually everyone, the probability of adoption is barely more than 50%.

Figure 15. Policy responsiveness
Based on 1,791 proposed policy changes over the years 1981-2002. Data source: Benjamin I. Page and Martin Gilens, Democracy in America?, University of Chicago Press, 2017, figure 3.2. The line is a linear regression line.

The difficulty in getting majority preferences enacted into policy is accentuated where there are two dominant ideologically-cohesive and ideologically-distinct political parties, each supported by roughly half of the electorate, one of which adopts an extreme oppositional stance. In such a context, a veto-point-heavy government structure may cross the line from caution to gridlock.31

Consistent with this worry, we’ve seen growing use of the filibuster during the era of party polarization, as figure 16 shows.

Figure 16. Use of the filibuster in the Senate
Number of cloture filings. Data source:, “Senate Action on Cloture Motions.” The line is a loess curve.

So has the federal government descended into legislative paralysis? One measure suggests yes. Figure 17 shows a steady decline in the number of laws passed by congress.

Figure 17. Number of laws passed by congress
Data sources: Tobin Grant, personal communication; GovTrack, “Bills by Final Status.” The line is a loess curve.

However, GovTrack, which compiles these data on the number of laws passed notes the following: “Since World War II (the earliest we have data), Congress has typically enacted 4-6 million words of new law in each two-year Congress. However, those words have been enacted in fewer but larger bills. Therefore, the generally decreasing number of bills enacted into law does not reflect less legislative work is occurring.”32 David Mayhew has compiled data on the number of “major” laws passed by congress each year. These are shown in figure 18. There is no indication of a decline.

Figure 18. Number of major laws passed by congress
Data source: David Mayhew, “Datasets and Materials: Divided We Govern”; The Economist, “Graphic Detail: American Lawmaking,” 2023. The line is a linear regression line.

So it may be that despite our extensive veto points and extensive party polarization, there’s been less gridlock than widely assumed.


Are there systemic biases in our political system that favor the election prospects of the Republicans or the Democrats? Demographic developments, especially the rising share of Latinos in the population, have tilted toward the Democrats. The same is true of cohort trends: each group of young Americans tends to be more liberal than the previous.33 Yet Republicans have fared better in elections since 1980 than they did in the previous half century, as figure 19 shows. Is that due to structural features of America’s democracy?

Figure 19. Partisan control of the presidency, the Senate, and the House of Representatives
Dots indicate Democratic control. Blank spaces indicate Republican control.

Since at least the 1950s, Americans with lower incomes have been more likely to vote Democratic while those with higher incomes have tended to vote Republican.34 The class skew in voting, with voter turnout lower among those with less education and incomes, has therefore tended to benefit Republicans.

Several features of our electoral system also tend to advantage the Republican Party. One is systemic: winner-take-all systems such as ours are less conducive to electoral success by left parties than proportional representation systems.35

At the moment, the Senate’s overrepresentation of small states favors the Republicans, because there are more conservative small states than liberal ones. Figure 20 offers one way to see this. It shows state population by the partisan leaning of each state. If we consider states in which a party got more than 53% of the two-party vote to be solidly partisan, there are 15 solidly Democratic states and 23 solidly Republican ones. In the solidly Republican states, there is one senator per 2 million people. In the solidly Democratic states, there is one senator per 4 million people.

Figure 20. Senate malapportionment favors the Republicans
The average population per senator for the 23 solidly Republican states (below 47% Democratic vote share in the 2012 presidential election) is 2.3 million. The average for the 15 solidly Democratic states (above 53% Democratic vote share) is 4.1 million. Data sources: Census Bureau; Wikipedia.

As noted above, Republicans in state government had more opportunity to gerrymander House districts following the 2010 census than did Democrats, and quite a few took advantage of that opportunity. This gave Republicans a boost in House elections, and in each election during the 2010s Republicans’ share of seats in the House of Representatives has exceeded their share of votes received. But as noted earlier, the growing use of independent commissions for redistricting may have reduced this problem.

The Republican Party has tended to be more committed than the Democrats to advancing the economic well-being of affluent individuals and business firms. Businesses and the rich have gotten a large and growing share of the country’s income since the late 1970s, and they have strengthened their political organization considerably. We might expect this to have produced a flood of money to Republicans. The Supreme Court’s 2010 Citizens United ruling prohibited restrictions on political campaign spending by organizations, such as firms and unions, opening the door even further to expenditures by outside groups on behalf of their preferred candidate or party.

Yet the evidence doesn’t support the notion that a money advantage has been the key reason behind recent Republican electoral success. Figure 21 shows campaign expenditures by and for Democrats and Republicans since 1998 (the earliest year for which reliable data are available). Republicans have not dominated here.

Figure 21. Campaign expenditures by and for Democrats and Republicans
Billions of inflation-adjusted dollars. Includes expenditures by candidates, parties, and outside groups. Data source: Center for Responsive Politics, “Cost of Election,”

It’s too soon to be able to render an informed judgment on the impact of the Citizens United decision, but the degree to which it altered the legal landscape is sometimes overstated.36 Before the super PACs (political action committees) and 501(c)(4)s that sprang up after Citizens United, individuals and corporations already could make unlimited donations to 527s. The only difference is that the new organizations are less constrained in naming the candidates they favor or oppose in advertisements running during the two months prior to the election.

Even if money totals do come to systematically favor Republicans at some point in the future, it’s unclear how much that will matter. There are diminishing returns to money in influencing election outcomes: when a lot is already being spent, additional amounts are likely to have limited impact.

Moreover, the chief determinant of the outcome of national elections is the health of the economy. Douglas Hibbs and Larry Bartels point out that presidential election outcomes can be predicted fairly well using just a single measure of economic performance: income growth in the months preceding the election.37 This is displayed in figure 22. On the vertical axis is the incumbent-party candidate’s popular vote margin. On the horizontal axis is the growth rate of real disposable personal income per capita in the middle two quarters (April through September) of the election year, adjusted for how long the incumbent party has been in office. This simple model does a very good job of predicting the vote outcome. Other models can predict even more accurately by including additional factors, but measures of economic performance are central in all of them.38

Figure 22. Income growth and presidential election outcomes
Vertical axis: incumbent-party candidate’s popular vote margin. Data source: Wikipedia. Horizontal axis: growth rate of real disposable personal income per capita in the second and third quarters of the election year, adjusted for incumbency (-1.29 for each consecutive term, beyond the first, that the incumbent party has held the White House). Data source: Bureau of Economic Analysis,, using National Income and Product Accounts (NIPA) data. This replicates and extends Larry Bartels’ chart in “Obama Toes the Line,” The Monkey Cage, 2013. For more detail, see Christopher Achen and Larry Bartels, Democracy for Realists, Princeton University Press, 2016, ch. 6. The correlation is +.91.

House and Senate elections are more idiosyncratic than presidential elections, yet the condition of the national economy tends to be a good predictor for them too.39 Incumbency also has a big effect. As figure 23 shows, a very large share of congressional incumbents who run for reelection win — 93% in the House and 81% in the Senate, on average, since 1950. That’s partly because they are able to raise more money than challengers, but it’s also because they develop name recognition, they can point to things they’ve accomplished for the district, and they have multiple ways of getting media exposure.40

Figure 23. Incumbent reelection success rate in congress
Share of those running for reelection who win. The lines are loess curves. Data sources: Norman Ornstein et al, Vital Statistics on Congress, Brookings Institution, tables 2.7 and 2.8; Center for Responsive Politics, “Reelection Rates Over the Years,”

Money plays a role in election outcomes, to be sure, but it appears to be no more important as a source of Republican advantage, and perhaps less so, than our winner-take-all elections, Senate malapportionment, House district gerrymandering, and the class skew in voter turnout.


While there is much to cheer in America’s political system, particularly its stability over time, there also is quite a bit to lament. Too few Americans vote, and participation in voting is heavily class-skewed, due in part to needlessly stiff registration and voting requirements. Effective translation of votes into elected representatives is impeded by key aspects of our electoral system — winner-take-all elections, Senate malapportionment, House of Representatives district gerrymandering, and the presidential electoral college. High-income Americans have much more influence on policy decisions than do those with lower incomes. And America’s multitude of governmental veto points may obstruct policy change.

However, the news is not all bad. New opportunities to vote early and by mail have increased voter turnout. As more states have turned electoral redistricting for the House over to independent commissions, gerrymandering has decreased. Inequality of political influence doesn’t appear to have increased since the late 1970s even though income and wealth inequality have risen sharply. And despite the potential for our veto-point-heavy political system to be paralyzed in the context of a polarized and evenly balanced electorate, passage of major legislation by the federal government hasn’t decreased.

  1. Robert A. Dahl, Democracy and Its Critics, Yale University Press, 1989; John Rawls, Justice as Fairness: A Restatement, Harvard University Press, 2001; Dahl, On Political Equality, Yale University Press, 2006; Joshua Cohen, “Money, Politics, Political Equality,” in Cohen, Philosophy, Politics, Democracy, Harvard University Press, 2009. 
  2. Jens Alber and Ulrich Kohler, “The Inequality of Electoral Participation in Europe and America and the Politically Integrative Functions of the Welfare State,” Social Science Research Center Berlin (WZB), 2008; Heather K. Gerken, “Make It Easy: The Case for Automatic Registration,” Democracy, 2013; Theodore J. Lowi, Benjamin Ginsberg, Kenneth A. Shepsle, and Stephen Ansolabehere, American Government: Power and Purpose, 12th edition, W.W. Norton, 2012, ch. 11. 
  3. Nick Corasaniti and Jim Rutenberg, “Republicans Pushed to Restrict Voting. Millions of Americans Pushed Back,” New York Times, 2020. 
  4. ACE Electoral Knowledge Network, “Electoral Systems,” 
  5. Frances E. Lee and Bruce I. Oppenheimer, Sizing Up the Senate: The Unequal Consequences of Equal Representation, University of Chicago Press, 1999; Arend Lijphart, Patterns of Democracy, Yale University Press, 1999; Robert Dahl, How Democratic Is the American Constitution?, 2nd edition, Yale University Press, 2003; Sanford Levinson, Our Undemocratic Constitution, Oxford University Press, 2006. 
  6. The 48 number includes two independents who caucus with the Democrats. 
  7. Lee and Oppenheimer, Sizing Up the Senate, table 4.6. 
  8. “2016 National House Popular Vote Tracker — Google Sheets,” retrieved Dec 6, 2016. 
  9. Seth E. Masket, Jonathan Winburn, and Gerald C. Wright, “The Gerrymanderers Are Coming! Legislative Redistricting Won’t Affect Competition or Polarization Much, No Matter Who Does It,” PS: Political Science and Politics, 2012; Jowei Chen and David Cottrell, “Evaluating Partisan Gains from Congressional Gerrymandering: Using Computer Simulations to Estimate the Effect of Gerrymandering in the U.S. House,” Electoral Studies, 2016; Jonathan Rodden, Why Cities Lose, Basic Books, 2019. 
  10. Robert Draper, “The League of Dangerous Mapmakers,” The Atlantic, 2012; Sam Wang, “The Great Gerrymander of 2012,” New York Times, 2013; Nicholas Goedert, “Gerrymandering or Geography? How Democrats Won the Popular Vote but Lost the Congress in 2012,” Research and Politics, 2014; Christopher Ingraham, “This Is Actually What America Would Look Like without Gerrymandering,” Washington Post: Wonkblog, 2016; Michael Wines, “Judges Find Wisconsin Redistricting Unfairly Favored Republicans,” New York Times, 2016. 
  11. Maine and Nebraska are exceptions. 
  12. Paul Burstein, American Public Opinion, Advocacy, and Policy in Congress: What the Public Wants and What it Gets, Cambridge University Press, 2014; Christopher Achen and Larry M. Bartels, Democracy for Realists, Princeton University Press, 2016. 
  13. Achen and Bartels, Democracy for Realists, ch. 2. 
  14. Arthur Lupia and John G. Matsusaka, “Direct Democracy: New Approaches to Old Questions,” Annual Review of Political Science, 2004. 
  15. David Broder, Democracy Derailed: Initiative Campaigns and the Power of Money, Harcourt, 2000; Achen and Bartels, Democracy for Realists, ch. 3. 
  16. For an optimistic take, see Benjamin I. Page and Robert Y. Shapiro, “Effects of Public Opinion on Policy,” American Political Science Review, 1983; Robert S. Erikson, Michael B. MacKuen, and James A. Stimson, The Macro Polity, Cambridge University Press, 2002. 
  17. Think of Don Regan, former CEO of Merrill Lynch, Treasury Secretary and Chief of Staff for President Reagan; Robert Rubin, former Chair of Goldman Sachs, Treasury Secretary for President Clinton; and quite a few members of Donald Trump’s cabinet. 
  18. Schlozman et al, The Unheavenly Chorus, especially ch. 5. See also Sidney Verba, Kay Lehman Schlozman, Henry E. Brady, Voice and Equality: Civic Voluntarism in American Politics, Harvard University Press, 1995. 
  19. E.E. Schattschneider, The Semisovereign People, Holt, Rinehart, and Winston, 1960; Thomas Ferguson and Joel Rogers, Right Turn: The Decline of the Democrats and the Future of American Politics, Hill and Wang, 1986; Jacob S. Hacker and Paul Pierson, Winner-Take-All Politics, Simon and Schuster, 2010. See also David Vogel, Fluctuating Fortunes: The Political Power of Business in America, Basic Books, 1989; Jill Quadagno, “Social Movements and State Transformation: Labor Unions and Racial Conflict in the War on Poverty,” American Sociological Review, 1992; Edwin Amenta, Neil Caren, and Sheera Joy Olasky, “Age for Leisure? Political Mediation and the Impact of the Pension Movement on U.S. Old Age Policy,” American Sociological Review, 2005; Isaac William Martin, The Permanent Tax Revolt: How the Property Tax Transformed American Politics, Stanford University Press, 2008. 
  20. Alexander Hicks, Social Democracy and Welfare Capitalism, Cornell University Press, 1999; Wolfgang Streeck and Lane Kenworthy, “Theories and Practices of Neo-Corporatism,” in Handbook of Political Sociology, edited by Thomas Janoski et al, Cambridge University Press, 2005; Detlef Jahn, “Changing of the Guard: Trends in Corporatist Arrangements in 42 Highly Industrialized Societies from 1960 to 2010,” Socio-Economic Review, 2014. 
  21. Larry Bartels, Unequal Democracy, 2nd edition, Princeton University Press, 2016; Martin Gilens, Affluence and Influence, Princeton University Press, 2012; Martin Gilens and Benjamin I. Page, “Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens,” Perspectives on Politics, 2014. 
  22. Vogel, Fluctuating Fortunes; Hacker and Pierson, Winner-Take-All Politics. 
  23. Stephen Ansolabehere, John de Figueiredo, and James M. Snyder Jr., “Why Is There So Little Money in U.S. Politics?,” Journal of Economic Perspectives, 2006; Frank R. Baumgartner, Jeffrey M. Berry, Marie Hojnacki, David C. Kimball, and Beth L. Leech, Lobbying and Public Policy, University of Chicago Press, 2009; Ezra Klein, “Our Corrupt Politics — It’s Not All Money,” New York Review of Books, 2012; Burstein, American Public Opinion, Advocacy, and Policy in Congress. 
  24. Frederick Solt, “Economic Inequality and Democratic Political Engagement,” American Journal of Political Science, 2008. 
  25. Kay Lehman Schlozman, Sidney Verba, and Henry E. Brady, The Unheavenly Chorus: Unequal Political Voice and the Broken Promise of American Democracy, Princeton University Press, 2012, ch. 6. 
  26. Lane Kenworthy, “Income Distribution,” The Good Society. 
  27. Hacker and Pierson, Winner-Take-All Politics. 
  28. Lane Kenworthy, “Capitalism, Inequality, and Democracy,” The Good Society. 
  29. James Madison, The Federalist Papers, 10 and 51. 
  30. Ezra Klein, “What Happens When Congress Fails to Do Its Job?,” Newsweek, 2010; Michael Tomasky, “The Specter Haunting the Senate,” New York Review of Books, 2010. 
  31. Lane Kenworthy, “Is America Too Polarized?,” The Good Society; Thomas E. Mann and Norman J. Ornstein, It’s Even Worse Than It Looks, Basic Books, 2012; Ornstein and Mann, “The Republicans Waged a Three-Decade War on Government. They Got Trump,” Vox, 2016. 
  32. GovTrack, “Bills by Final Status.” 
  33. John B. Judis and Ruy Teixeira, The Emerging Democratic Majority, Scribner, 2002; Ruy Teixeira, The Optimistic Leftist, St. Martin’s Press, 2017. 
  34. Bartels, Unequal Democracy; Andrew Gelman, David Park, Boris Shor, and Jeronimo Cortina, Red State, Blue State, Rich State, Poor State: Why Americans Vote the Way They Do, 2nd edition, Princeton University Press, 2009. 
  35. Torben Iversen and David Soskice, “Electoral Institutions and the Politics of Coalitions: Why Some Democracies Redistribute More Than Others,” American Political Science Review, 2006. 
  36. Matt Bai, “How Much Has Citizens United Changed the Political Game?,” New York Times, 2012. 
  37. Douglas Hibbs, “The Bread and Peace Model Applied to the 2008 US Presidential Election,”, 2009; Larry Bartels, “Obama Toes the Line,” The Monkey Cage, 2013. 
  38. Larry Bartels and John Zaller, “Presidential Vote Models: A Recount,” PS: Political Science and Politics, 2001; Ray Fair, Predicting Presidential Elections and Other Things, 2nd edition, Stanford University Press, 2012; Nate Silver, “Measuring the Effect of the Economy on Elections,” FiveThirtyEight, 2012; John Sides and Lynn Vavreck, The Gamble: Choice and Chance in the 2012 Presidential Election, Princeton University Press, 2013. 
  39. Douglas Hibbs, “The Partisan Division of House Seats in 2012: Implications of the ‘Bread and Incumbency’ Model,”, 2012; Eric McGhee, “Forecasting House Elections,” The Monkey Cage, 2012. 
  40. Gary Jacobson, The Politics of Congressional Elections, 8th edition, Pearson, 2012; Lee Drutman, “Why Money Still Matters,” The Monkey Cage, 2012.